As the size and complexity of network simulations accessible to computational neuroscience grows, new avenues open for research into extracellularly recorded electric signals. Biophysically detailed simulations permit the identification of the biological origins of the different components of recorded signals, the evaluation of signal sensitivity to different anatomical, physiological, and geometric factors, and selection of recording parameters to maximize the signal information content. Simultaneously, virtual extracellular signals produced by these networks may become important metrics for neuro-simulation validation. To enable efficient calculation of extracellular signals from large neural network simulations, we have developed BlueRecording, a pipeline consisting of standalone Python code, along with extensions to the Neurodamus simulation control application, the CoreNEURON computation engine, and the SONATA data format, to permit online calculation of such signals. In particular, we implement a general form of the reciprocity theorem, which is capable of handling non-dipolar current sources, such as may be found in long axons and recordings close to the current source, as well as complex tissue anatomy, dielectric heterogeneity, and electrode geometries. To our knowledge, this is the first application of this generalized (i.e., non-dipolar) reciprocity-based approach to simulate EEG recordings. We use these tools to calculate extracellular signals from an in silico model of the rat somatosensory cortex and hippocampus and to study signal contribution differences between regions and cell types.

Advances in computational capabilities and large volumes of experimental data have established computer simulations of brain tissue models as an important pillar in modern neuroscience. Alongside, a variety of domain specific languages (DSLs) have been developed to succinctly express properties of these models, ensure their portability to different platforms, and provide an abstraction that allows scientists to work in their comfort zone of mathematical equations, delegating concerns about performance optimizations to downstream compilers. One of the popular DSLs in modern neuroscience is the NEURON MODeling Language (NMODL). Until now, its compilation process has been split into first transpiling NMODL to C++ and then using a C++ toolchain to emit the efficient machine code. This approach has several drawbacks including the reliance on different programming models to target heterogeneous hardware, maintainability of multiple compiler back-ends and the lack of flexibility to use the domain information for C++ code optimization. To overcome these limitations, we present MOD2IR, a new open-source code generation pipeline for NMODL. MOD2IR leverages the LLVM toolchain to target multiple CPU and GPU hardware platforms. Generating LLVM IR allows the vector extensions of modern CPU architectures to be targeted directly, producing optimized SIMD code. Additionally, this gives MOD2IR significant potential for further optimizations based on the domain information available when LLVM IR code is generated. We present experiments showing that MOD2IR is able to produce on-par execution performance using a single compiler back-end implementation compared to code generated via state-of-the-art C++ compilers, and can even surpass them by up to 1.26×. Moreover, MOD2IR supports JIT-execution of NMODL, yielding both efficient code and an on-the-fly execution workflow.

Ultraliser is a neuroscience-specific software framework capable of creating accurate and biologically realistic 3D models of complex neuroscientific structures at intracellular (e.g. mitochondria and endoplasmic reticula), cellular (e.g. neurons and glia) and even multicellular scales of resolution (e.g. cerebral vasculature and minicolumns). Resulting models are exported as triangulated surface meshes and annotated volumes for multiple applications in in silico neuroscience, allowing scalable supercomputer simulations that can unravel intricate cellular structure-function relationships. Ultraliser implements a high-performance and unconditionally robust voxelization engine adapted to create optimized watertight surface meshes and annotated voxel grids from arbitrary non-watertight triangular soups, digitized morphological skeletons or binary volumetric masks. The framework represents a major leap forward in simulation-based neuroscience, making it possible to employ high-resolution 3D structural models for quantification of surface areas and volumes, which are of the utmost importance for cellular and system simulations. The power of Ultraliser is demonstrated with several use cases in which hundreds of models are created for potential application in diverse types of simulations. Ultraliser is publicly released under the GNU GPL3 license on GitHub (BlueBrain/Ultraliser). SIGNIFICANCE: There is crystal clear evidence on the impact of cell shape on its signaling mechanisms. Structural models can therefore be insightful to realize the function; the more realistic the structure can be, the further we get insights into the function. Creating realistic structural models from existing ones is challenging, particularly when needed for detailed subcellular simulations. We present Ultraliser, a neuroscience-dedicated framework capable of building these structural models with realistic and detailed cellular geometries that can be used for simulations.

MOTIVATION: Astrocytes, the most abundant glial cells in the mammalian brain, have an instrumental role in developing neuronal circuits. They contribute to the physical structuring of the brain, modulating synaptic activity and maintaining the blood-brain barrier in addition to other significant aspects that impact brain function. Biophysically, detailed astrocytic models are key to unraveling their functional mechanisms via molecular simulations at microscopic scales. Detailed, and complete, biological reconstructions of astrocytic cells are sparse. Nonetheless, data-driven digital reconstruction of astroglial morphologies that are statistically identical to biological counterparts are becoming available. We use those synthetic morphologies to generate astrocytic meshes with realistic geometries, making it possible to perform these simulations. RESULTS: We present an unconditionally robust method capable of reconstructing high fidelity polygonal meshes of astroglial cells from algorithmically-synthesized morphologies. Our method uses implicit surfaces, or metaballs, to skin the different structural components of astrocytes and then blend them in a seamless fashion. We also provide an end-to-end pipeline to produce optimized two- and three-dimensional meshes for visual analytics and simulations, respectively. The performance of our pipeline has been assessed with a group of 5000 astroglial morphologies and the geometric metrics of the resulting meshes are evaluated. The usability of the meshes is then demonstrated with different use cases. AVAILABILITY AND IMPLEMENTATION: Our metaball skinning algorithm is implemented in Blender 2.82 relying on its Python API (Application Programming Interface). To make it accessible to computational biologists and neuroscientists, the implementation has been integrated into NeuroMorphoVis, an open source and domain specific package that is primarily designed for neuronal morphology visualization and meshing. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Accurate morphological models of brain vasculature are key to modeling and simulating cerebral blood flow in realistic vascular networks. This in silico approach is fundamental to revealing the principles of neurovascular coupling. Validating those vascular morphologies entails performing certain visual analysis tasks that cannot be accomplished with generic visualization frameworks. This limitation has a substantial impact on the accuracy of the vascular models employed in the simulation. RESULTS: We present VessMorphoVis, an integrated suite of toolboxes for interactive visualization and analysis of vast brain vascular networks represented by morphological graphs segmented originally from imaging or microscopy stacks. Our workflow leverages the outstanding potentials of Blender, aiming to establish an integrated, extensible and domain-specific framework capable of interactive visualization, analysis, repair, high-fidelity meshing and high-quality rendering of vascular morphologies. Based on the initial feedback of the users, we anticipate that our framework will be an essential component in vascular modeling and simulation in the future, filling a gap that is at present largely unfulfilled. AVAILABILITY AND IMPLEMENTATION: VessMorphoVis is freely available under the GNU public license on Github at https://github.com/BlueBrain/VessMorphoVis. The morphology analysis, visualization, meshing and rendering modules are implemented as an add-on for Blender 2.8 based on its Python API (application programming interface). The add-on functionality is made available to users through an intuitive graphical user interface, as well as through exhaustive configuration files calling the API via a feature-rich command line interface running Blender in background mode. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The evaluation of small degree polynomials is critical for the computation of elementary functions. It has been extensively studied and is well documented. In this article, we evaluate existing methods for polynomial evaluation on superscalar architecture. In addition, we have completed this work with a factorization method, which is surprisingly neglected in the literature. This work focuses on out-of-order Intel processors, amongst others, of which computational units are available. Moreover, we applied our work on the elementary function e x that requires, in the current implementation, an evaluation of a polynomial of degree 10 for a satisfying precision and performance. Our results show that the factorization scheme is the fastest in benchmarks, and that latency and throughput are intrinsically dependent on each other on superscalar architecture.

Exposing parallelism in scientific applications has become a core requirement for efficiently running on modern distributed multicore SIMD compute architectures. The granularity of parallelism that can be attained is a key determinant for the achievable acceleration and time to solution. Motivated by a scientific use case that requires the simulation of long spans of time - the study of plasticity and learning in detailed models of brain tissue - we present a strategy that exposes and exploits multicore and SIMD micro-parallelism from unrolling flow dependencies and concurrent outputs in a large system of coupled ordinary differential equations (ODEs). An implementation of a parallel simulator is presented, running on the HPX runtime system for the ParalleX execution model, providing dynamic taskscheduling and asynchronous execution. The implementation was tested on different architectures using a previously published brain tissue model. Benchmark of single neurons on a single compute node present a speed-up of circa 4-7× when compared with the state of the art Single Instruction Multiple Data (SIMD) implementation and 13-40× over its Single Instruction Single Data (SISD) counterpart. Large scale benchmarks suggest almost ideal strong scaling and a speed-up of 2-8× on a distributed architecture of 128 Cray X6 compute nodes.

Motivation: From image stacks to computational models, processing digital representations of neuronal morphologies is essential to neuroscientific research. Workflows involve various techniques and tools, leading in certain cases to convoluted and fragmented pipelines. The existence of an integrated, extensible and free framework for processing, analysis and visualization of those morphologies is a challenge that is still largely unfulfilled. Results: We present NeuroMorphoVis, an interactive, extensible and cross-platform framework for building, visualizing and analyzing digital reconstructions of neuronal morphology skeletons extracted from microscopy stacks. Our framework is capable of detecting and repairing tracing artifacts, allowing the generation of high fidelity surface meshes and high resolution volumetric models for simulation and in silico imaging studies. The applicability of NeuroMorphoVis is demonstrated with two case studies. The first simulates the construction of three-dimensional profiles of neuronal somata and the other highlights how the framework is leveraged to create volumetric models of neuronal circuits for simulating different types of in vitro imaging experiments. Availability and implementation: The source code and documentation are freely available on https://github.com/BlueBrain/NeuroMorphoVis under the GNU public license. The morphological analysis, visualization and surface meshing are implemented as an extensible Python API (Application Programming Interface) based on Blender, and the volume reconstruction and analysis code is written in C++ and parallelized using OpenMP. The framework features are accessible from a user-friendly GUI (Graphical User Interface) and a rich CLI (Command Line Interface). Supplementary information: Supplementary data are available at Bioinformatics online.

Every neuron is part of a network, exerting its function by transforming multiple spatiotemporal synaptic input patterns into a single spiking output. This function is specified by the particular shape and passive electrical properties of the neuronal membrane, and the composition and spatial distribution of ion channels across its processes. For a variety of physiological or pathological reasons, the intrinsic input/output function may change during a neuron's lifetime. This process results in high variability in the peak specific conductance of ion channels in individual neurons. The mechanisms responsible for this variability are not well understood, although there are clear indications from experiments and modeling that degeneracy and correlation among multiple channels may be involved. Here, we studied this issue in biophysical models of hippocampal CA1 pyramidal neurons and interneurons. Using a unified data-driven simulation workflow and starting from a set of experimental recordings and morphological reconstructions obtained from rats, we built and analyzed several ensembles of morphologically and biophysically accurate single cell models with intrinsic electrophysiological properties consistent with experimental findings. The results suggest that the set of conductances expressed in any given hippocampal neuron may be considered as belonging to two groups: one subset is responsible for the major characteristics of the firing behavior in each population and the other is responsible for a robust degeneracy. Analysis of the model neurons suggests several experimentally testable predictions related to the combination and relative proportion of the different conductances that should be expressed on the membrane of different types of neurons for them to fulfill their role in the hippocampus circuitry.

BACKGROUND: We present a visualization pipeline capable of accurate rendering of highly scattering fluorescent neocortical neuronal models. The pipeline is mainly developed to serve the computational neurobiology community. It allows the scientists to visualize the results of their virtual experiments that are performed in computer simulations, or in silico. The impact of the presented pipeline opens novel avenues for assisting the neuroscientists to build biologically accurate models of the brain. These models result from computer simulations of physical experiments that use fluorescence imaging to understand the structural and functional aspects of the brain. Due to the limited capabilities of the current visualization workflows to handle fluorescent volumetric datasets, we propose a physically-based optical model that can accurately simulate light interaction with fluorescent-tagged scattering media based on the basic principles of geometric optics and Monte Carlo path tracing. We also develop an automated and efficient framework for generating dense fluorescent tissue blocks from a neocortical column model that is composed of approximately 31000 neurons. RESULTS: that is reconstructed from the somatosensory cortex of juvenile rat. The fluorescence optical model is qualitatively analyzed and validated against experimental emission spectra of different fluorescent dyes from the Alexa Fluor family. CONCLUSION: We discussed a scientific visualization pipeline for creating images of synthetic neocortical neuronal models that are tagged virtually with fluorescent labels on a physically-plausible basis. The pipeline is applied to analyze and validate simulation data generated from neuroscientific in silico experiments.

BACKGROUND: We present a software workflow capable of building large scale, highly detailed and realistic volumetric models of neocortical circuits from the morphological skeletons of their digitally reconstructed neurons. The limitations of the existing approaches for creating those models are explained, and then, a multi-stage pipeline is discussed to overcome those limitations. Starting from the neuronal morphologies, we create smooth piecewise watertight polygonal models that can be efficiently utilized to synthesize continuous and plausible volumetric models of the neurons with solid voxelization. The somata of the neurons are reconstructed on a physically-plausible basis relying on the physics engine in Blender. RESULTS: Our pipeline is applied to create 55 exemplar neurons representing the various morphological types that are reconstructed from the somatsensory cortex of a juvenile rat. The pipeline is then used to reconstruct a volumetric slice of a cortical circuit model that contains ∼210,000 neurons. The applicability of our pipeline to create highly realistic volumetric models of neocortical circuits is demonstrated with an in silico imaging experiment that simulates tissue visualization with brightfield microscopy. The results were evaluated with a group of domain experts to address their demands and also to extend the workflow based on their feedback. CONCLUSION: A systematic workflow is presented to create large scale synthetic tissue models of the neocortical circuitry. This workflow is fundamental to enlarge the scale of in silico neuroscientific optical experiments from several tens of cubic micrometers to a few cubic millimeters. AMS SUBJECT CLASSIFICATION: Modelling and Simulation.

Scientific workflows are often composed of compute-intensive simulations and data-intensive analysis and visualization, both equally important for productivity. High-performance computers run the compute-intensive phases efficiently, but data-intensive processing is still getting less attention. Dense non-volatile memory integrated into super-computers can help address this problem. In addition to density, it offers significantly finer-grained I/O than disk-based I/O systems. We present a way to exploit the fundamental capabilities of Storage-Class Memories (SCM), such as Flash, by using scalable key-value (KV) I/O methods instead of traditional file I/O calls commonly used in HPC systems. Our objective is to enable higher performance for on-line and near-line storage for analysis and visualization of very high resolution, but correspondingly transient, simulation results. In this paper, we describe 1) the adaptation of a scalable key-value store to a BlueGene/Q system with integrated Flash memory, 2) a novel key-value aggregation module which implements coalesced, function-shipped calls between the clients and the servers, and 3) the refactoring of a scientific workflow to use application-relevant keys for fine-grained data subsets. The resulting implementation is analogous to function-shipping of POSIX I/O calls but shows an order of magnitude increase in read and a factor 2.5x increase in write IOPS performance (11 million read IOPS, 2.5 million write IOPS from 4096 compute nodes) when compared to a classical file system on the same system. It represents an innovative approach for the integration of SCM within an HPC system at scale.

Large multiscale neuronal network simulations are of increasing value as more big data are gathered about brain wiring and organization under the auspices of a current major research initiative, such as Brain Research through Advancing Innovative Neurotechnologies. The development of these models requires new simulation technologies. We describe here the current use of the NEURON simulator with message passing interface (MPI) for simulation in the domain of moderately large networks on commonly available high-performance computers (HPCs). We discuss the basic layout of such simulations, including the methods of simulation setup, the run-time spike-passing paradigm, and postsimulation data storage and data management approaches. Using the Neuroscience Gateway, a portal for computational neuroscience that provides access to large HPCs, we benchmark simulations of neuronal networks of different sizes (500-100,000 cells), and using different numbers of nodes (1-256). We compare three types of networks, composed of either Izhikevich integrate-and-fire neurons (I&F), single-compartment Hodgkin-Huxley (HH) cells, or a hybrid network with half of each. Results show simulation run time increased approximately linearly with network size and decreased almost linearly with the number of nodes. Networks with I&F neurons were faster than HH networks, although differences were small since all tested cells were point neurons with a single compartment.

BACKGROUND: We present a physically-based computational model of the light sheet fluorescence microscope (LSFM). Based on Monte Carlo ray tracing and geometric optics, our method simulates the operational aspects and image formation process of the LSFM. This simulated, in silico LSFM creates synthetic images of digital fluorescent specimens that can resemble those generated by a real LSFM, as opposed to established visualization methods producing visually-plausible images. We also propose an accurate fluorescence rendering model which takes into account the intrinsic characteristics of fluorescent dyes to simulate the light interaction with fluorescent biological specimen. RESULTS: We demonstrate first results of our visualization pipeline to a simplified brain tissue model reconstructed from the somatosensory cortex of a young rat. The modeling aspects of the LSFM units are qualitatively analysed, and the results of the fluorescence model were quantitatively validated against the fluorescence brightness equation and characteristic emission spectra of different fluorescent dyes. AMS SUBJECT CLASSIFICATION: Modelling and simulation.

The spatial arrangement of Ca2+ channels and vesicles remains unknown for most CNS synapses, despite of the crucial importance of this geometrical parameter for the Ca2+ control of transmitter release. At a large model synapse, the calyx of Held, transmitter release is controlled by several Ca2+ channels in a "domain overlap" mode, at least in young animals. To study the geometrical constraints of Ca2+ channel placement in domain overlap control of release, we used stochastic MCell modelling, at active zones for which the position of docked vesicles was derived from electron microscopy (EM). We found that random placement of Ca2+ channels was unable to produce high slope values between release and presynaptic Ca2+ entry, a hallmark of domain overlap, and yielded excessively large release probabilities. The simple assumption that Ca2+ channels can be located anywhere at active zones, except below a critical distance of ~ 30 nm away from docked vesicles ("exclusion zone"), rescued high slope values and low release probabilities. Alternatively, high slope values can also be obtained by placing all Ca2+ channels into a single supercluster, which however results in significantly higher heterogeneity of release probabilities. We also show experimentally that high slope values, and the sensitivity to the slow Ca2+ chelator EGTA-AM, are maintained with developmental maturation of the calyx synapse. Taken together, domain overlap control of release represents a highly organized active zone architecture in which Ca2+ channels must obey a certain distance to docked vesicles. Furthermore, domain overlap can be employed by near-mature, fast-releasing synapses.

Scientists in many disciplines use spatial mesh models to study physical phenomena. Simulating natural phenomena by changing meshes over time helps to better understand the phenomena. The higher the precision of the mesh models, the more insight do the scientists gain and they thus continuously increase the detail of the meshes and build them as detailed as their instruments and the simulation hardware allow. In the process, the data volume also increases, slowing down the execution of spatial range queries needed to monitor the simulation considerably. Indexing speeds up range query execution, but the overhead to maintain the indexes is considerable because almost the entire mesh changes unpredictably at every simulation step. Using a simple linear scan, on the other hand, requires accessing the entire mesh and the performance deteriorates as the size of the dataset grows. In this paper we propose OCTOPUS, a strategy for executing range queries on mesh datasets that change unpredictably during simulations. In OCTOPUS we use the key insight that the mesh surface along with the mesh connectivity is sufficient to retrieve accurate query results efficiently. With this novel query execution strategy, OCTOPUS minimizes index maintenance cost and reduces query execution time considerably. Our experiments show that OCTOPUS achieves a speedup between 7.3 and 9.2× compared to the state of the art and that it scales better with increasing mesh dataset size and detail.

Neuroscientists increasingly use computational tools in building and simulating models of the brain. The amounts of data involved in these simulations are immense and efficiently managing this data is key. One particular problem in analyzing this data is the scalable execution of range queries on spatial models of the brain. Known indexing approaches do not perform well even on today's small models which represent a small fraction of the brain, containing only few millions of densely packed spatial elements. The problem of current approaches is that with the increasing level of detail in the models, also the overlap in the tree structure increases, ultimately slowing down query execution. The neuroscientists' need to work with bigger and more detailed (denser) models thus motivates us to develop a new indexing approach. To this end we develop FLAT, a scalable indexing approach for dense data sets. We base the development of FLAT on the key observation that current approaches suffer from overlap in case of dense data sets. We hence design FLAT as an approach with two phases, each independent of density. In the first phase it uses a traditional spatial index to retrieve an initial object efficiently. In the second phase it traverses the initial object's neighborhood to retrieve the remaining query result. Our experimental results show that FLAT not only outperforms R-Tree variants from a factor of two up to eight but that it also achieves independence from data set size and density.

Today's scientists are quickly moving from in vitro to in silico experimentation: they no longer analyze natural phenomena in a petri dish, but instead they build models and simulate them. Managing and analyzing the massive amounts of data involved in simulations is a major task. Yet, they lack the tools to efficiently work with data of this size. One problem many scientists share is the analysis of the massive spatial models they build. For several types of analysis they need to interactively follow the structures in the spatial model, e.g., the arterial tree, neuron fibers, etc., and issue range queries along the way. Each query takes long to execute, and the total time for executing a sequence of queries significantly delays data analysis. Prefetching the spatial data reduces the response time considerably, but known approaches do not prefetch with high accuracy. We develop SCOUT, a structure-aware method for prefetching data along interactive spatial query sequences. SCOUT uses an approximate graph model of the structures involved in past queries and attempts to identify what particular structure the user follows. Our experiments with neuro-science data show that SCOUT prefetches with an accuracy from 71% to 92%, which translates to a speedup of 4x-15x. SCOUT also improves the prefetching accuracy on datasets from other scientific domains, such as medicine and biology.

We present a process to automatically generate three-dimensional mesh representations of the complex, arborized cell membrane surface of cortical neurons (the principal information processing cells of the brain) from nonuniform morphological measurements. Starting from manually sampled morphological points (3D points and diameters) from neurons in a brain slice preparation, we construct a polygonal mesh representation that realistically represents the continuous membrane surface, closely matching the original experimental data. A mapping between the original morphological points and the newly generated mesh enables simulations of electrophysiolgical activity to be visualized on this new membrane representation. We compare the new mesh representation with the state of the art and present a series of use cases and applications of this technique to visualize simulations of single neurons and networks of multiple neurons.

The rich dynamical nature of neurons poses major conceptual and technical challenges for unraveling their nonlinear membrane properties. Traditionally, various current waveforms have been injected at the soma to probe neuron dynamics, but the rationale for selecting specific stimuli has never been rigorously justified. The present experimental and theoretical study proposes a novel framework, inspired by learning theory, for objectively selecting the stimuli that best unravel the neuron's dynamics. The efficacy of stimuli is assessed in terms of their ability to constrain the parameter space of biophysically detailed conductance-based models that faithfully replicate the neuron's dynamics as attested by their ability to generalize well to the neuron's response to novel experimental stimuli. We used this framework to evaluate a variety of stimuli in different types of cortical neurons, ages and animals. Despite their simplicity, a set of stimuli consisting of step and ramp current pulses outperforms synaptic-like noisy stimuli in revealing the dynamics of these neurons. The general framework that we propose paves a new way for defining, evaluating and standardizing effective electrical probing of neurons and will thus lay the foundation for a much deeper understanding of the electrical nature of these highly sophisticated and non-linear devices and of the neuronal networks that they compose.

The thick-tufted layer 5b pyramidal cell extends its dendritic tree to all six layers of the mammalian neocortex and serves as a major building block for the cortical column. L5b pyramidal cells have been the subject of extensive experimental and modeling studies, yet conductance-based models of these cells that faithfully reproduce both their perisomatic Na(+)-spiking behavior as well as key dendritic active properties, including Ca(2+) spikes and back-propagating action potentials, are still lacking. Based on a large body of experimental recordings from both the soma and dendrites of L5b pyramidal cells in adult rats, we characterized key features of the somatic and dendritic firing and quantified their statistics. We used these features to constrain the density of a set of ion channels over the soma and dendritic surface via multi-objective optimization with an evolutionary algorithm, thus generating a set of detailed conductance-based models that faithfully replicate the back-propagating action potential activated Ca(2+) spike firing and the perisomatic firing response to current steps, as well as the experimental variability of the properties. Furthermore, we show a useful way to analyze model parameters with our sets of models, which enabled us to identify some of the mechanisms responsible for the dynamic properties of L5b pyramidal cells as well as mechanisms that are sensitive to morphological changes. This automated framework can be used to develop a database of faithful models for other neuron types. The models we present provide several experimentally-testable predictions and can serve as a powerful tool for theoretical investigations of the contribution of single-cell dynamics to network activity and its computational capabilities.

This work presents an incremental training approach that allows for the use of simple evolutionary algorithms to efficiently train the weights of fast mixed-signal hardware neural networks. The training strategy is tested on a set of common classification benchmark problems (E.coli, yeast, thyroid and wine) and the results are comparable to those of other investigations. They demonstrate that the used hardware neural network architecture can successfully be trained to solve real-world problems. It is also shown that the presented training strategy is suited for use in conjunction with a specialized coprocessor that speeds up evolutionary algorithms by performing the genetic operations within a configurable logic. With the ability to integrate the coprocessor into the training strategy, an important prerequisite is fulfilled to exploit the speed of the neural hardware during training.

Computation without stable states is a computing paradigm dif- ferent from Turing's and has been demonstrated for various types of simulated neural networks. This publication transfers this to a hardware implemented neural network. Results of a software im- plementation are reproduced showing that the performance peaks when the network exhibits dynamics at the edge of chaos. The liquid computing approach seems well suited for operating analog computing devices such as the used VLSI neural network.

This paper presents experimental results obtained during the training of an analog hardware neural network for the prediction of cellular localization sites of proteins in yeast and E.coli. The synaptic weights of the network are optimized by an evolutionary chip-in-the-loop algorithm. The results provide a first demonstration of the applicability of the presented neural network architecture to real-world problems that require the ability to generalize and to handle quasi-continuous multi-bit inputs.